检索结果-内蒙古大学图书馆

37th Australasian Joint Conference on artificial intelligence, AJCAI 2024

作者： Rimas, Mohamad Chen, Qi Zhang, Mengjie Centre for Data Science and Artificial Intelligence and School of Engineering and Computer Science Victoria University of Wellington Wellington New Zealand

ISBN: (纸本)9789819603503

Pruning is a major research field in neural networks, enhancing their efficiency and generalization. The field of pruning approaches in genetic programming (GP) is continually evolving, with researchers actively exploring new techniques and approaches to optimise the performance of GP models. This research introduces a novel pruning algorithm for Genetic Programming-based Symbolic Regression (GPSR). The proposed method employs a weighting mechanism to identify and filter out unimportant subtrees in each generation. To achieve this, the method arranges all subtrees linearly and assigns weights to each subtree and terminal. It then uses Ordinary Least Squares (OLS) to optimize these weights, enabling the identification of unimportant subtrees and terminals for effective pruning. The algorithm’s effectiveness was evaluated using ten regression datasets, including high-dimensional and complex feature sets. Furthermore, comparisons were made with two other algorithms to evaluate its performance. The results indicate that the proposed approach not only achieves better learning and generalisation performance but also generates smaller trees compared to standard GP, thereby improving interpretability. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Genetic Programming Ordinary Least Squares Pruning Regression Subtrees

来源：评论

学校读者我要写书评

暂无评论

Knowledge As Not Only Justified True Beliefs in Vision

引用

IEEE/CAA Journal of Automatica Sinica 2025年第2期12卷 297-299页

作者： Wenbo Zheng Fei-Yue Wang IEEE School of Computer Science and Artificial Intelligence Wuhan University of Technology Sanya Science and Education Innovation Park of Wuhan University of Technology the State Key Laboratory for Management and Control of Complex Systems Institute of Automation Chinese Academy of Sciences

COMPUTATIONAL knowledge vision [1] is emphasized as a novel perspective or field in this paper. It first proposes the visual hierarchy and its connection to knowledge, stating that knowledge is a justified true belief. To further the previous research, we concisely summarize our recent works and suggest a new direction that knowledge is also a thought framework in vision.

关键词： Computational modeling Visualization artificial intelligence data models Visual perception Visual databases Training data

来源：评论

学校读者我要写书评

暂无评论

Beyond Labels and Topics: Discovering Causal Relationships in Neural Topic Modeling 24

Beyond Labels and Topics: Discovering Causal Relationships i...

引用

33rd ACM Web Conference, WWW 2024

作者： Tang, Yi-Kun Huang, Heyan Shi, Xuewen Mao, Xian-Ling School of Computer Science and Technology Beijing Institute of Technology Beijing China School of Data Science and Artificial Intelligence Dongbei University of Finance and Economics Liaoning Dalian China

ISBN: (纸本)9798400701719

Topic models that can take advantage of labels are broadly used in identifying interpretable topics from textual data. However, existing topic models tend to merely view labels as names of topic clusters or as categories of texts, thereby neglecting the potential causal relationships between supervised information and latent topics, as well as within these elements themselves. In this paper, we focus on uncovering possible causal relationships both between and within the supervised information and latent topics to better understand the mechanisms behind the emergence of the topics and the labels. To this end, we propose Causal Relationship-Aware Neural Topic Model (CRNTM), a novel neural topic model that can automatically uncover interpretable causal relationships between and within supervised information and latent topics, while concurrently discovering high-quality topics. In CRNTM, both supervised information and latent topics are treated as nodes, with the causal relationships represented as directed edges in a Directed Acyclic Graph (DAG). A Structural Causal Model (SCM) is employed to model the DAG. Experiments are conducted on three public corpora with different types of labels. Experimental results show that the discovered causal relationships are both reliable and interpretable, and the learned topics are of high quality comparing with eight start-of-the-art topic model baselines. © 2024 ACM.

关键词： Directed graphs

来源：评论

学校读者我要写书评

暂无评论

Diabetes Prediction: Machine Learning Models Enhanced by Harmony Search Optimization 1

Diabetes Prediction: Machine Learning Models Enhanced by Har...

引用

1st International Conference on Advanced Computing and Emerging Technologies, ACET 2024

作者： Vennu, Mohith Kokkirala, Ajay Roa Saketh Reddy, Donthi Dwivedi, Anshu Kumar Sr University School of Computer Science and Artificial Intelligence Warangal India Sr University School of Computer Science and Artificial Intelligence Warangal India Bennett University School of Computer Science Engineering and Technology Greater Noida India

ISBN: (纸本)9798350367720

The Harmony Search (HS) algorithm has been used in many applications, including engineering, excavation, epidemiology, object recognition, stock price prediction, genetic analysis, and structural design optimization, all derived from the improvisation process in musicians. The study covers the application of the HS algorithm for a predictive model of diabetes with its parameter tuning and optimization process of the model. In performing the research, a dataset will be used that contains data of information regarding demographic data, clinical metrics, and lifestyle attributes of people with and without diabetes. The most important features such as age, BMI, blood pressure, and glucose level will be treated as input variables. The HS algorithm will optimize the model parameters to develop the best model for the process of prediction of diabetes. The research has four parts: preparing, choosing features, training models, HS-driven upgrade, and checking results. The best model with K-NN beat all basic models in accuracy, sensitivity, and specificity. This shows good risk evaluation and early diabetes detection. This article contributes new contributions to emerging areas of diabetes prediction through novel avenues in model optimization. Specifically, this shows the efficacy of the HS algorithm, which improves the performance of predictive models. In that sense, this improves the ability to do an accurate and reliable risk evaluation of diabetes in clinical applications. © 2024 IEEE.

关键词： Clinical research

来源：评论

学校读者我要写书评

暂无评论

Collaborative Offloading Method for Digital Twin Empowered Cloud Edge Computing on Internet of Vehicles

引用

Tsinghua science and technology 2023年第3期28卷 433-451页

作者： Linjie Gu Mengmeng Cui Linkun Xu Xiaolong Xu Changwang School of Honors Nanjing University of Information Science and TechnologyNanjing 210044China School of Computer and Software Nanjing University of Information Science and TechnologyNanjing 21044China School of Artificial Intelligence Nanjing University of Information and TechnologyNanjing 210044China

Digital twinning and edge computing are attractive solutions to support computing-intensive and servicesensitive Internet of Vehicles *** of the existing Internet of Vehicles service offloading solutions only consider edge–cloud collaboration,but the collaboration between small cell eNodeB(SCeNB)should not be *** delays far lower than offloading tasks to the cloud can be obtained through reasonable collaborative computing between *** proposed framework realizes and maintains the simulation of collaboration between SCeNB nodes by constructing a digital twin that maintains SCeNB nodes in the central controller,thereby realizing user task offloading positions,sub-channel allocation,and computing resource *** an algorithm named AUC-AC is proposed,based on the dominant actor–critic network and the auction *** order to obtain a better command of global information,the convolutional block attention mechanism(CBAM)is used in the digital twin of each SCeNB node to observe its environment and learn *** results show that our experimental scheme is better than several baseline algorithms in terms of service delay.

关键词： digital twin cloud-edge computing reinforcement learning actor-critic network

来源：评论

学校读者我要写书评

暂无评论

Learning from Heterogeneity: A Dynamic Learning Framework for Hypergraphs

IEEE Transactions on Artificial Intelligence

引用

IEEE Transactions on artificial intelligence 2025年第6期6卷 1513-1528页

作者： Zhang, Tiehua Liu, Yuze Shen, Zhishu Ma, Xingjun Qi, Peng Ding, Zhijun Jin, Jiong Tongji University School of Computer Science and Technology Shanghai201804 China Ant Group Shanghai200122 China Wuhan University of Technology School of Computer Science and Artificial Intelligence Wuhan430070 China Fudan University School of Computer Science Shanghai200438 China Tongji University Department of Control Science and Engineering Shanghai201804 China Swinburne University of Technology School of Science Computing and Engineering Technologies Melbourne3122 Australia

Graph neural network (GNN) has gained increasing popularity in recent years owing to its capability and flexibility in modeling complex graph structure data. Among all graph learning methods, hypergraph learning is a technique for exploring the implicit higher-order correlations when training the embedding space of the graph. In this article, we propose a hypergraph learning framework named learning from heterogeneity (LFH) that is capable of dynamic hyperedge construction and attentive embedding update utilizing the heterogeneity attributes of the graph. Specifically, in our framework, the high-quality features are first generated by the pairwise fusion strategy that utilizes explicit graph structure information when generating initial node embedding. Afterward, a hypergraph is constructed through the dynamic grouping of implicit hyperedges, followed by the type-specific hypergraph learning process. To evaluate the effectiveness of our proposed framework, we conduct comprehensive experiments on several popular datasets with twelve state-of-the-art models on both node classification and link prediction tasks, which fall into categories of homogeneous pairwise graph learning, heterogeneous pairwise graph learning, and hypergraph learning. The experimental results demonstrate a significant performance gain (an average of 12.9% in node classification and 12.8% in link prediction) compared with recent state-of-the-art methods. © 2020 IEEE.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

A Closer Look at the CLS Token for Cross-Domain Few-Shot Learning 38

A Closer Look at the CLS Token for Cross-Domain Few-Shot Lea...

引用

38th Conference on Neural Information Processing Systems, NeurIPS 2024

作者： Zou, Yixiong Yi, Shuai Li, Yuhua Li, Ruixuan School of Computer Science and Technology Huazhong University of Science and Technology China School of Artificial Intelligence and Automation Huazhong University of Science and Technology China

Vision Transformer (ViT) has shown great power in learning from large-scale datasets. However, collecting sufficient data for expert knowledge is always difficult. To handle this problem, Cross-Domain Few-Shot Learning (CDFSL) has been proposed to transfer the source-domain knowledge learned from sufficient data to target domains where only scarce data is available. In this paper, we find an intriguing phenomenon neglected by previous works for the CDFSL task based on ViT: leaving the CLS token to random initialization, instead of loading source-domain trained parameters, could consistently improve target-domain performance. We then delve into this phenomenon for an interpretation. We find the CLS token naturally absorbs domain information due to the inherent structure of the ViT, which is represented as the low-frequency component in the Fourier frequency space of images. Based on this phenomenon and interpretation, we further propose a method for the CDFSL task to decouple the domain information in the CLS token during the source-domain training, and adapt the CLS token on the target domain for efficient few-shot learning. Extensive experiments on four benchmarks validate our rationale and state-of-the-art performance. Our codes are available at https://***/Zoilsen/CLS_Token_CDFSL. © 2024 Neural information processing systems foundation. All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Bird Nest Detection in Railway Catenaries Using a Coarse-to-Fine Strategy

引用

Journal of Engineering science and technology Review 2024年第4期17卷 99-108页

作者： Zheng, Xin Ruan, Maoliang Zhang, Xingyu Lu, Junbo Luo, Xuegang School of Artificial Intelligence and Bigdata Sichuan University of Arts and Science DaZhou635000 China School of Mathematics and Computer Science Panzhihua University Panzhihua617000 China

Catenaries installed above railway lines are frequently exposed to the outdoor environment, making them prone to accumulating foreign objects, such as bird nests. These represent a primary risk to operational safety. Currently, the main method of catenary bird nest detection involves the manual analysis of video images to identify and mark nests. This method is inefficient and susceptible to human error due to fatigue, leading to missed detections and failure to identify and remove nests in a timely and accurate manner. This study addressed the challenge of intelligently detecting bird nests in high-speed railway catenaries by introducing an advanced analysis method that integrates initial detection with precise identification using deep-learning technology. Initially, the YOLO-v5 network was used to detect suspected bird nest areas from catenary monitoring images. This was followed by precise identification using the EfficientNet-B4 network. Furthermore, a comprehensive dataset comprising 15,000 images under various imaging conditions, including normal, blurred imaging, foggy conditions, partial exposure, and partial obstruction of bird nests, was constructed. The experimental results on the catenary monitoring image dataset demonstrate that the YOLO-v5 network achieves an accuracy of 0.9269 and a recall of 0.9210 on the test set. The performance is further enhanced through precise recognition using the EfficientNet-B4 network, which achieves an accuracy of 0.9479 and a recall of 0.9180. This research not only surpasses existing methods in performance but also demonstrates significant potential for application in detecting bird nests on railway catenaries. Moving beyond traditional manual identification and inspection, this study leverages deep learning to achieve the precise and rapid detection of bird nests, thereby enhancing the speed and accuracy of inspections. This advancement promotes the automation of catenary inspections and ensures the operational safety

关键词： Railroads

来源：评论

学校读者我要写书评

暂无评论

LF-DETR: Lightweight Detection Network for Floating Surface Garbage Based on Adaptive Channel Shuffling 22

LF-DETR: Lightweight Detection Network for Floating Surface ...

引用

22nd IEEE International Symposium on Parallel and Distributed Processing with Applications, ISPA 2024

作者： Xu, ShouKun Gong, Yan Yang, Gaochao Li, Ning School of Software Changzhou University School of Computer Science and Artificial Intelligence Aliyun School of Big Data Changzhou China

ISBN: (纸本)9798331509712

To address the issues of low accuracy and efficiency in water surface target detection caused by uneven lighting and water ripples, this study proposes a channel modeling-based DETR-like model. Starting from efficient partial convolutions and inverted residual blocks, the concept of dynamic sparsity is introduced, and a Flexible Masked Residual Block (FMRB) for a lightweight backbone is designed. Based on the design principle of channel grouping, an adaptive channel shuffling strategy is derived. Our model achieved an mAP of 79.8% on the FGDD dataset and 61.7% on the Pascal VOC2007 dataset. Compared to the original model, the FLOPs of our LF-DETR model are reduced by 24.9%, down to 42.8G, and the number of parameters is reduced by 29.4%. Extensive experimental results demonstrate that our model exhibits an excellent balance between efficiency and accuracy across multiple benchmarks, highlighting its great potential in practical applications. © 2024 IEEE.

关键词： Object detection

来源：评论

学校读者我要写书评

暂无评论

A Novel Method for Microbial Detection in Wastewater Treatment

引用

Journal of Engineering science and technology Review 2024年第5期17卷 153-162页

作者： Zheng, Xin Tang, Chengyu Lv, Junrui Luo, Xuegang School of Artificial Intelligence and Bigdata Sichuan University of Arts and Science DaZhou635000 China School of Mathematics and Computer Science Panzhihua University Panzhihua617000 China

In sewage disposal processes, real-time and accurate monitoring of activated sludge microorganism types and quantities can facilitate prompt adjustments to process parameters, thereby mitigating lag effects inherent in sewage treatments. Traditionally, microscopic examinations of activated sludge microorganisms have relied on manual observation and counting by personnel to assess microorganism types and quantities under varying operational conditions. This method is inefficient and produces results lacking accuracy and consistency. To address the challenges associated with activated sludge microorganism detection, an intelligent analysis method underpinned by deep learning technology was proposed in this study. The methodology encompassed three primary stages: pre-processing of activated sludge microorganism micrographs, feature extraction from image blocks using a dedicated module, and determination of microorganism types and locations via a microbial information analysis module. To support this study, an activated sludge microorganism image dataset was constructed, comprising 2,000 images of four microorganism species: Epistylis, Nematoda, Lecane, and Arcella. Experimental results demonstrate that, the proposed method achives a mean absolute error of 5.86 and a mean squared error of 12.43 on the test set, surpassing the performance of six existing comparison methods. This exceptional performance underscores the method's significant potential for application in activated sludge microorganism detection. This study marks a substantial advancement over conventional manual detection techniques by leveraging deep learning to accurately identify activated sludge microorganisms. The proposed method enhances the speed, accuracy, and consistency of microorganism detection, thereby contributing to increased automation in water quality monitoring during sewage treatment. Ultimately, this study promotes intelligent efficiency in sewage treatments and provides a strong foundati

关键词： Wastewater treatment

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：